Drew and Arash went to MIT, where their everyday computing environment was UNIX systems with home directories on AFS. For those who haven't used it, AFS is a networked filesystem sort of like NFS, except with a better authentication model and the ability to support failover on the server side.

AFS, NFS, Coda, 9P, WebDAV, SMB/CIFS, FTP, Windows roaming profiles, AFP, etc. were all around and (to varying levels) built into OSes well before Dropbox. Dropbox was a fundamentally different model: files stay on your local machine and get synchronized, instead of living on the network. The entire idea that live synchronization is the correct approach is almost entirely due to Dropbox's success; any OS designer before 2007 would have said "Yes, we have something as basic as a networked filesystem."

If you're objecting to the fact that it takes a commercial venture to explore and develop these sorts of ideas, I too dislike that we're in that place. But it's true. We haven't achieved a way to develop standard computer infrastructure without capitalism, yes; but we've certainly achieved meaningful technology that did not previously exist that's actually working for people.

(And, also, almost all of those protocols I mentioned were developed by profitable businesses like Sun or Transarc or Apple or whomever, and only made non-proprietary at great effort. The ones that started off with public specs, like WebDAV and FTP, were the worst.)

> If you're objecting to the fact that it takes a commercial venture to explore and develop these sorts of ideas, I too dislike that we're in that place. But it's true. We haven't achieved a way to develop standard computer infrastructure without capitalism

While private business certainly is essential, let's not go too far. I can think of ... maybe just one or two pieces of standard computer infrastructure not developed by private business or for profit. Need I list them?

Doesn't seem like a correct characterization from my perspective. File distributions was "hot" from the early 00s until cloud everything of recent years. I think Dropbox is/was clever but not surprising, because that was the sort of things a lot of people were doing (or speculating about) at the time. Especially since most things were based on files and not web like today. The main reason it didn't develop wasn't so much lack of insight as that the Internet become increasingly hostile towards stand alone solutions, including the increase in phones and other different clients.

I definitely think that part of Dropbox bet was that the Internet wouldn't develop much in the future. A bet they aren't alone in making.

"Dropbox was a fundamentally different model"

You mention a bunch of filesystems (some of which with much better scalability than Dropbox), but omit Rsync. Rsync works with any of these filesystems.

The most successful data distribution systems which predate these are from the '90s and before Usenet (esp volume with the binaries) and scene (or FXP). However, the former is mostly public data and the latter is inherently insecure.

Dropbox was nothing special. It was just the first popular one who had a GUI on top of Rsync, and who sold data storage within a package. Its essentially a lucky shot.

WebDAV works great for client-server, but it has a disadvantage: locking works terrible. OwnCloud/NextCloud uses WebDAV, btw.

Dropbox is a lot more than a GUI on top of rsync. Even purely from an engineering standpoint (ignoring product & design) that's incorrect.

You might enjoy this talk: https://www.youtube.com/watch?v=PE4gwstWhmc

Live-synchronization became a thing when the use-model of the internet shifted from always-connected to sometimes-connected. I don't think that nobody ever thought of this synchronization model before. I mean, it's pretty close to rsync or git.

Doesn't AFS keep local copies of files too?

That’s similar to the line Steve Jobs took when negotiating with them. Hard to tell if that was just nasty Steve coming out during a negotiation that wasn’t going favourably, or if that’s what he thought. Given that Steve wanted Dropbox, I’d gues he was just posturing.

https://www.businessinsider.com.au/drew-houston-dropbox-stev...

As long as NAT is the default mostly everywhere we'll be stuck with 3rd parties to interact between users, if only to poke hole in those NATs. You can fire up a local HTTP server on your machine very easily if you want to share a file but it's no use if your IP is not reachable.

Of course there's more to Dropbox than that, you also have 3rd party "cloud" storage which is an important feature, especially if you want your file to remain available to 3rd parties when you shutdown your machine.

It pretty much is a standard part of the OS. iCloud, OneDrive, etc.

It's the "non-proprietary" part you're not ever going to see. Someone has to pay for storage.

I think that GP's point was that a standard a la HTML or RSS should exist and be embraced by the OS. Something like Syncthing, I guess (just naming one of the first hits I got from searching, I haven't really kept up with what is being tried in this space).

https://en.wikipedia.org/wiki/Syncthing

Don't be so sure -- IPFS provides a possible path. We will see if it works.

Someone, somewhere has to pay for the disk.

Yes IPFS the collective internet prices and earns for contributing disk.

It's the protocol that's non-proprietary, not the service. IP is non-proprietary but you still have to pay for servers.

Why would non-proprietary preclude paying for storage?

Because each of those is trying to profit from it, not just pay for the storage. Thus, branding, and thus, they make them private and proprietary. Is it right? No.

> they have turned something that should have been a commodity into a profitable business

Commodity businesses are universally profitable businesses for the survivors. The companies don't stick around otherwise.

Convenience stores, typical pizza places, liquor stores, tons of grocery store products, most office supplies, generic clothing, power adapters, most physical digital storage, brick & cement companies, the 58000 self storage facilities, the world is filled with profitable commodity products and businesses (and they're often very large).

The secret is that it's not basic.

If it were basic, then it would be standard as you say. Therefore it must be a lot more complicated.

It's basic in the sense that git and rsync are basic: they are standard tools that everybody expects nowadays. I'm not saying the tech is trivial.

He achieved three commas.